48176_en-US

liliconcom

Predictive Model

Hi. I'm trying to build a model using several different variables(X1 to X5) having as Y an economic output. The customer wants to know how does that Y changes as well as the X's when modifing some constants  (input prices) which are not part of the model but part of the formula that calculates the Y. Is there any way within a model in JMP that this can be done other than going to the formula and modifing those values? Thaks for your time

Re: Predictive Model

One way is to make new columns to hold your parameters then use the profiler. Remember you can re-size the x-axis on the graph for the parameter inside the profiler.

 

An example:

 

Names default to here( 1 );

dt = Open( "$Sample_data/Big class.jmp" );

//Column with the initial parameter value
dt << New Column( "X1",
 Numeric,
 "Continuous",
 Format( "Best", 12 ),
 Formula( 1 )
);

//Column with model formula
dt << New Column( "My Formula",
 Numeric,
 "Continuous",
 Format( "Best", 12 ),
 Formula( Parameter( {b0 = 1}, :height + :weight + :X1 ) )
);

//Profiler now includes the parameter
dt << Profiler(
 Y( :My Formula ),
 Profiler(
  1,
  Term Value(
   height( 61.25, Lock( 0 ), Show( 1 ) ),
   weight( 120, Lock( 0 ), Show( 1 ) ),
   X1( 0.77, Min( -10 ), Max( 10 ), Lock( 0 ), Show( 1 ) )
  )
 ),
 SendToReport(
  Dispatch(
   {"Prediction Profiler"},
   "3",
   ScaleBox,
   {Min( -10 ), Max( 10 ), Inc( 2 ), Minor Ticks( 0 )}
  )
 )
);